Recognizing a raised object on the basis of perspective images

1. A method for recognizing a raised object in an environmental region of a motor vehicle, comprising the steps: capturing a first image from a first camera and a second image from a second camera with a perspective that is different from the first camera, wherein the environmental region is imaged by the first image and the second image each at least partly, transforming the first image and the second image into a common reference system according to a respective predetermined transformation instruction, forming gradients over pixel values of pixels along respective gradient lines in the transformed first image and the transformed second image, summing gradients along several parallel summation lines to form a respective gradient sum for the transformed first image and the transformed second image, wherein the summation lines and the gradient lines are perpendicular to each other and maxima of the gradient sum represent boundaries of the raised object, determining a first pair of the maxima of the gradient sum in the transformed first image and a second pair of the maxima of the gradient sum in the transformed second image, wherein the pair according to a predetermined criterion correspond with each other, recognizing the raised object, if at least one distance between the maxima of the first pair in the transformed first image deviates from a distance of the maxima of the corresponding second pair in the transformed second image.

2. The method according to claim 1 , wherein the first image and the second image are transformed into a bird's eye perspective as the common reference system.

3. The method according to claim 1 , wherein as the raised object, a border stone is recognized.

4. The method according to claim 1 , wherein the raised object is assigned a height.

5. The method according to claim 1 , wherein an overlapping region of the environmental region is imaged by the first image and the second image, and the raised object is recognized at least partly within the overlapping region.

6. The method according to claim 5 , wherein a further overlapping region of the environmental region is imaged by a further first image from a further first camera and a further second image from a further second camera and the raised object is recognized by interpolation between the overlapping region and the further overlapping region.

7. The method according to claim 1 , wherein the raised object is recognized by extrapolation of the first image and/or the second image in a partial portion of the environmental region, which is imaged by maximally the first or the second image.

8. The method according to claim 1 , wherein for determining the gradient sum to start with a main orientation direction of the raised object is recognized and the gradient lines are aligned perpendicular to the main orientation direction of the raised object.

9. A driver assistance device for recognizing a raised object in an environmental region of a motor vehicle, which is configured to perform the method according to claim 1 .

10. A motor vehicle with a driver assistance device for recognizing a raised object in an environmental region of the motor vehicle, configured for performing the method according to claim 1 , comprising: a first camera capturing a first image and a second camera with a perspective that is different from the first camera for capturing a second image, wherein the cameras are configured to image the environmental region by the first image and the second image each at least partially, a geometry unit for transforming the first image and the second image into a common reference system according to the respective predetermined transformation instruction, a calculation unit for forming a gradient over pixels along respective gradient lines in the transformed first image and the transformed second image, for summing the gradients along several parallel summation lines to form a respective gradient sum for the transformed first image and the transformed second image, wherein the summation lines and the gradient lines are perpendicular to each other and maxima of the gradient sums represent boundaries of the raised object, for determining a first pair of the maxima of the gradient sum in the transformed first image and a second pair of maxima of the gradient sum in the transformed second image, wherein the pairs according to a predetermined criterion correspond with each other, and for recognizing the raised object, if at least a distance between the maxima of the first pair in the transformed first image deviates from a distance of the maxima of the corresponding second pair in the transformed second image.

11. A method for recognizing a raised object in an environmental region of a motor vehicle, comprising the steps: capturing a first image from a first camera and a second image from a second camera with a perspective that is different from the first camera, wherein the environmental region is imaged by the first image and the second image each at least partly, transforming the first image and the second image into a common reference system according to a respective predetermined transformation instruction, forming gradients over pixel values of pixels along respective gradient lines in the transformed first image and the transformed second image, summing gradients along several parallel summation lines to form a respective gradient sum for the transformed first image and the transformed second image, wherein the summation lines and the gradient lines are perpendicular to each other and maxima of the gradient sum represent boundaries of the raised object, forming a progression of difference from the gradient sum of the transformed first image and the gradient sum of the transformed second image, determining the raised object, when the progression of difference and/or the integral over the progression of difference exceeds a predetermined threshold value.

12. A motor vehicle with a driver assistance device for recognizing a raised object in an environmental region of the motor vehicle, configured for performing the method according to claim 11 , comprising: a first camera capturing a first image and a second camera with a perspective that is different from the first camera for capturing a second image, wherein the cameras are configured to image the environmental region by the first image and the second image each at least partially, a geometry unit for transforming the first image and the second image into a common reference system according to the respective predetermined transformation instruction, a calculation unit for forming a gradient over pixels along respective gradient lines in the transformed first image and the transformed second image, for summing the gradients along several parallel summation lines to form a respective gradient sum for the transformed first image and the transformed second image, wherein the summation lines and the gradient lines are perpendicular to each other and maxima of the gradient sums represent boundaries of the raised object, for forming a progression of difference from the gradient sum of the transformed first image and the gradient sum of the transformed second image and for determining the raised object, when the progression of difference and/or the integral over the progression of difference exceeds a predetermined threshold value.